Construction of a Multi-Parametric Ultrasonographic Nomogram for Precise Assessment of Papillary Breast Lesions.

Abstract

To analyze the multi-parametric ultrasonographic (MpUS) features of atypical/malignant papillary lesions of the breast with clinical information, identify independent risk factors, and construct a nomogram to improve the diagnostic accuracy. This retrospective study analyzed consecutively hospitalized patients diagnosed with pathologically confirmed papillary breast lesions from January 2017 to June 2023. Preoperative sonographic exams, including gray-scale ultrasound (G-US), color Doppler flow imaging (CDFI), and contrast-enhanced ultrasound (CEUS), were conducted. Sonographic scans were retrospectively reviewed alongside clinicopathological data. Binary logistic regression identified independent risk factors for screening atypical/malignant papillary lesions. The receiver operating characteristic curve evaluated the predictive accuracy of these lesions, resulting in the development of a nomogram for assessing risk. The study involved 176 female patients with breast papillary lesions, identifying key predictors for atypical or malignant outcomes: age 57 or order, US diameter ≥13.95 mm, resistive index ≥0.70, enlarged enhancement on CEUS, and contrast agent retention, with respective odds ratios of 6.279, 8.078, 9.246, 9.401, and 5.047. The integrated use of G-US, CDFI, and CEUS in the MpUS approach offered higher diagnostic accuracy (AUC: 0.966) than G-US or CDFI alone (0.869/0.918). CEUS particularly enhanced prediction for non-mass-like lesions, with a positive predictive value of 83.3%. A nomogram incorporating MpUS and patient age achieved an AUC of 0.956 for predicting atypical or malignant papillary lesions. MpUS imaging is highly effective for predicting malignant breast papillary lesions, especially considering patient age. The nomogram offers an intuitive framework for assessing malignant risk in these lesions. Ultrasound excels in identifying papillary lesions, and integrating diverse data and multi-parametric imaging enhances malignant risk evaluation. This study establishes a predictive risk model using the nomogram method, demonstrating heightened diagnostic efficacy in breast papillary lesions.